Today’s policymakers have an opportunity to implement plans that will guide the efficient usage of AVs and rider choices that may affect us for generations. …

Ironically, the efficiency of AVs has long been touted as a solution to traffic, but new research is beginning to suggest that AVs will, in fact, generate more of it. …

By eliminating most of the hassles of driving, such as parking and lost productivity time, AVs will induce not only more trips, but longer ones. Additionally, AVs waiting to pick up new riders will add “deadheading” miles. For traffic, the only thing worse than a single-occupant vehicle is a zero-occupant vehicle. Placed all together, this suggests they will almost certainly increase vehicle-miles traveled, energy use, and emissions. These impacts might be locked in by further sprawl and other shifts toward less efficient land-use patterns. …

To avoid the worst of these traffic scenarios, policy needs to be deployed with an eye towards minimizing the added miles and the demand for situations involving zero-occupant vehicles. …

… policies should always seek to encourage AVs that move more people in fewer vehicles. While the driverless technologies make point-to-point drop-offs possible, the realities of cities and highways means that they simply cannot accommodate one AV per person. …

The deployment and pricing models offered by automotive and tech companies should be structured to make shared AVs, not personal AVs, the model of choice. …

… policymakers should seek to create pricing policies in anticipation of the traffic-inducing effects of personal AVs. The program might be created in escalating prices, as to disincentivize the least efficient choices. A VMT fee would discourage longer trips in general, while a higher single-occupant fee would encourage AV riders to share rides. Lastly, a zero-occupant fee, addressing the miles added by AVs circling between pick-ups or headed home to park, would warrant the highest fee. …

The national dialogue around AV policy is a unique chance to rethink how we prioritize our transportation systems and the incentives within it. A century ago, when the internal combustion engine automobile began to proliferate, cities missed this opportunity to guide how they affected communities.


Full story from Mobility Lab here.


  1. One effect of automated vehicles will be that senior citizens now banned from driving due to health concerns will potentially have another means of transportation available to them.

  2. Isn’t this somewhat similar to the Uber effect in Manhattan? The addition of Uber a few years ago has actually increased congestion some 8%

  3. Needs a major rethink of parking policies too.
    Cars won’t deadhead home, they’re more likely to deadhead to deadhead to the closest neighborhood with available on street parking. So the cars coming from Richmond into downtown will then park in Strathcona.
    I can also picture AV’s taking up 2hr zone parking spaces and then automatically moving to another 2hr space nearby when time is running out. This defeats the current intent to make these spaces convenient for local shoppers,etc.

    1. Yup. Parking has to be made more expensive. Like squatting, using public space you don’t own for an extended period for free has to end. https://pricetags.wordpress.com/2016/03/07/free-parking-is-like-squatting/
      AVs with zero occupants are many years off into the future. Many many. They make for interesting PhDs and grants from governments, though. We don’t even have Uber, so add a decade just for Vancouver. Likely not before 2040 here. Check back in 25 years due to all the legal, technical, human and infrastructure challenges involved. I was told speech recognition will be here “soon” 30 years ago, but when I tell my car ” phone Dave” or ” turn up temperature two degrees” it very frequently still says ” pardon?”.

  4. This issue is nothing new given that a transit bus with no passenger and just the driver is the same thing. That driver does not need to travel the route (i.e. unproductive trip) and is effectively invisible from an occupant perspective. I have collected and analysed transit ridership data and profiles and you rarely have a full bus load over the entire route, so not every bus is efficiently used. Many on even peak times can have just a few passenger km’s over the entire route. The calculation of transit efficiency from this perspective is actually a basic homework assignment question, but one that I don’t see being utilized by many transit planners because the stats sometimes can show gross transit inefficiency and therefore does not help the business case of expansion (which is actually decided on many other factors as well). Yes there are very effective routes, but these are unfortunately more akin to unicorns than your usual pony.
    And then of course deadheading is effectively a waste in terms of productivity (but operationally necessary). However, with AVs, their service needs may be much reduced, as well as having reduced deadheading, if at all. I have been working on concepts of roving service AVs that provide services to ride hailing AVs (yes robots servicing robots). Overall you have more variables to play with than you had in the past (bus depot locations were somewhat fixed) and therefore higher potential for optimization.
    Remember that the definition of an occupant assumes the trip the occupant is making is a productive one with a purpose beyond just driving around people for pay. So on that you can add to the list taxis and any other ride-hailing service.
    Rather than drivers driving around aimlessly looking for passengers, or on routes with no demand, you have connected vehicles working as a team and going into hibernation mode when not needed. You are not going to get nearly the same km’s of “zero-occupant” travel as you do currently in “zero-occupant” driven cars.
    In fact with all things equal, AVs would still be more efficient given there is one less human required to be transported–the driver. And depending on the driver’s weight that’s a lot of energy waste that can be reduced (a person can comprise of 3-5% of the total vehicular weight, and this would increase if lighter electric motors and batteries are produced, so savings are significant) .
    The research community is already all over this (for a good decade now?) and solutions are being proposed which are not different than many current logistical problems that are solved each day. We may not be able to optimize the logistics 100% but 90-95% optimal is feasible with current computer power and that could mean much less vehicles existing on our roads. And then if we run into computing limits, we can use quantum annealing algorithms to provide quick and almost complete solutions (and then you are using the mysterious powers of quantum teleportation which is an interesting combination to be had with surface transportation)
    It sounds as unlikely and crazy as going from rotary dial land-lines to quad-core smart phones that tell you if your cholesterol is high.
    As the private sector will be taking over in a big way, what we need to do (and really only can do) is start considering the consequences of this from a socio-economic perspective because AV’s have the potential to be more environmentally friendly, cheaper per km, and much safer than our current transportation setup. And so the business case is increasingly being set.

  5. Clark, I think there is a bit of counterpoint to the notion that “the private sector will be talking over in a big way.” Influencing certainly, but taking over?
    This is where the metrics of urbanism and geometry could temper this idea quite a bit. There is no more energy efficient, affordable and better mode of mobility in cities than high-capacity, frequent transit running in straight lines serving dense, compact urban areas, short of walking.
    Jarrett Walker dives into this topic quite a bit, and is especially critical of Uber and Lyft from the perspective of the geometry of urban mobility. I don’t think he has reached the point of illumination that was expressed in the recent work of James Arbib and Tony Seba yet (see link below), though with cheap, driverless EVs evolving essentially into huge driverless taxi fleets, human nature still has the capacity to disrupt the efficiencies of machines. A fleet of vehicle cleaning robots will no doubt be required in places like the entertainment district on weekend nights.

    1. Geometry is also a social problem. Along with the huge number of injuries and deaths, this is my main objection to motordom: not the environmental impact or economic inefficiency, but the dissolution of communities and the privatization of public streets and spaces.
      A life oriented around cars is a life lived at home and at work, with little or no public life in between: a life lived in homogeneous echo chambers, in other words, where we are seldom confronted with people unlike ourselves. As envisioned, cybercars threaten to make the situation worse, ensuring that at every moment we have the opportunity to step into our own little physical bubble (where we will doubtless disappear into the even more tenuous social and mental bubbles of our screens).

      1. Looking it up now, apparently it’s an old European Commission initiative: “Le Cybercar est un véhicule routier entièrement automatisé, au moins sur certaines infrastructures, sous contrôle d’un système de gestion, et pouvant transporter des personnes ou des marchandises.”
        Cybercar makes for clunky slang. Being electric, they will presumably hum, but Hummer is taken. The Google image results for cybercar look like bugs. So I’m thinking humbugs.

    2. Alex, i’m not surely advocating for a private sector takeover but I’m just echoing what I see as signs even stronger now, with trends obvious two decades ago. A key characteristic of TransLink’s design was to future-proof the system (the governance model was developed in the mid-90’s) but unfortunately that governance vehicle has been steered in the wrong direction.
      Having taught and researched on intelligent transportation systems (ITS) led me to realize that government was too slow to control the technology side of transportation (and I worked in all levels of government most of my career) and so you have what we see now with Uber and the boom of private-sector autonomous R&D. This was evident at early conferences back in 2000 where automobile manufacturers were right up there in ITS R&D. Any public-sector efforts that stood out were being lured into the private sector. As much as there are lots of public sector and academic publications on ITS and autonomous tech, the bulk of the innovative stuff are being developed by the private sector and not published (even I have stopped publishing on certain topics lest I give away all my intellectual property).
      I even recall about a decade ago we had an SFU discussion on this and I was on the panel. Some people thought I was nuts whereas I was simply describing technological change that was increasing exponentially, much like human population (and yes it will flatten out to an S curve but it may also bump up again if there is truly a technological singularity).
      On a side note, Jarrett Walker was a contractor at TransLink about the time I was project manager for the planning of what is now the Evergreen Line (early 2000’s). Funny thing is that I never heard him say one word ever and to me he was just this tall guy in the background helping the transit planners. Much of what is core to his thesis is very similar to what our director at the time, Clive Rock (one of the most innovative transport planners in Canada I would say), and I brought back from our trip visiting rapid transit lines and manufacturers in France: that the system for the Northeast sector needed to be at a “human scale” at level and not a massive system that requires lofty heights or subterranean portals, and we presented this concept in numerous occasions. The “human” and aesthetic emphasis was front and center for one of the manufacturers (Translohr) and we used it as a focus of system design concept. In fact the b/c revealed this system (a rubber-tyred tram) was the most cost-effective and had the most appropriate capacity for this area (you also didn’t need a tunnel). It was based on a 10 option multiple account evaluation with the reports being essentially wipe out of existence coincidentally when the business case was “improved” by the Province and we now have SkyTrain. But I digress…

    3. Clark, I lived in Coquitlam at the time, and I remember the BRT option when I attended the open house. As it was presented on paper, it did look like the best value for money. But I ranked it low when (as I recall) there was an opportunity to pick priorities as part of the consultation.
      First, even if it delivered what was promised, it appeared inferior to Skytrain or LRT. More bang for the buck, yes: but still, less bang.
      Second, experience. Growing up with Ottawa’s bus service (which after moving to Vancouver I realized was actually pretty good) there was no question in my mind that buses were vastly inferior to trains. They are not timely, they are not pleasant, their routes are not legible, their routes and schedules often get changed, and above all – the most important feature, especially when service is inconsistent – they are not frequent.
      Third, it was risky. According to the documents it was new technology with few existing implementations. So it might deliver as promised… or it might not. As a technologist, I have learned not to put undue trust in novelty.
      Fourth, it appeared to be a political attempt to cut costs by delivering something that looked flashy but would, in fact, be bog-standard bus service. This looked like the option designed to appeal to voters who did not plan to use the service anyway.
      From my point of view at the time, the choice was between rolling the dice and hoping that an unproven technology delivered, or paying more for a known and reliable solution.

      1. Geoff, we may have met at the open house as I was at pretty much all of them. Note that the BRT (bus rapid transit) option was the “best bus” option and was mainly used as a comparator or baseline relative to the “rapid transit” options. The rubber-tyred option was called the “GLT” or “gulded-light transit” and is essentially the same as an LRT other than rubber tires and the lack of rails (a central rail was for guiding the vehicles).
        As you know we went through a fairly wide-range of consultation from open houses, telephone surveys, focus groups, and even a “community leaders” panel who spent quite a bit of time reviewing all the work and information, with some going to France to further see the systems live. They actually voted for the GLT as the best system.
        I agree SkyTrain is the superior technology, and it is reflected by its cost. However it is ironic that people did emphasize the point (and it is valid) that GLT is a new technology, and it is therefore risky. I’m sure the same was said in the early 1980’s before SkyTrain was chosen. In the end, technological progress requires some level of risk.
        At any rate, my job was to ensure a thorough and fair process and evaluation. I was selected to be the pm for the project not because I had expertise in rapid transit, but the fact I did not and therefore I was not biased to any particular technology. The assurance that I do not bias the project keep it objective was reinforced to me by Board members, which was very refreshing as we started on a clean slate and I had no clue as to the outcome (nor should I care as my worry was to ensure a fair process).
        I also tried to take the opportunity to develop a standardized multiple-account evaluation framework for “mega projects” such that the same ruler is used to measure options for a particular project, or across the region between projects. If you noticed, this process (and phasing of technology and alignments) was used in other recent projects, as well as the Ottawa rapid transit project as they asked me for a copy of my reports and used the framework for their evaluations.
        So going back on topic, I believe we need to apply the same/similar evaluation process on any project, including autonomous technologies. A standardized and solution agnostic approach provides for a more fair, open, and eventually sustainable solution that takes on a regional and long-term purview, rather than a local, piece-meal and short-term approach.

Leave a Reply

Your email address will not be published. Required fields are marked *